JADE-GS: Joint Alternating Deblurring Guided by Events in 3D Gaussian Splatting
Computer Vision and Pattern Recognition
Summary
The authors explain that when cameras move quickly, images get blurry and hard to use for making 3D models. They use special event cameras that capture fast motion to fix this, but combining different methods to restore sharp images is tricky because each has weaknesses. Their system, JADE-GS, smartly mixes these methods and connects 2D image restoration with 3D modeling in a loop, which helps correct errors and improve results. They tested it on both fake and real data, finding it makes sharper and more accurate images and 3D models quickly on normal computers.
Authors
Haoyu Fu, Jiafeng Huang, Yuchen Wang, Shengjie Zhao
Abstract
When a camera moves fast during exposure, blur destroys the intra-exposure motion a 3D model needs to recover the sharp scene, while event cameras capture exactly this signal at microsecond resolution. Turning them into reliable 3D supervision faces two obstacles. First, the two restoration priors fail in opposite ways: physics-based event-integration priors preserve edges but accumulate drift; learned networks recover texture but distort boundaries. Second, existing pipelines run in one direction only, so raw event noise or the biases of fixed 2D pseudo-labels pass uncorrected into the geometry. JADE-GS addresses both: a pixel-adaptive routing gate fuses the complementary priors, and the resulting 2D restorer is coupled to a 3D Gaussian Splatting student in a bidirectional loop, where detached, multi-view-consistent renders and a physics-based reblurring constraint regularize the restorer, turning a fixed preprocessor into a geometry-aware predictor. Across synthetic and real benchmarks, JADE-GS attains the best perceptual quality, leading LPIPS and CLIP-IQA on both benchmarks with competitive PSNR and SSIM, and trainsin about one hour under 5 GB on a single consumer GPU while preserving real-time rendering.